Professional Documents
Culture Documents
Zietz2003 PDF
Zietz2003 PDF
Abstract
Copper in drinking water has been associated with Non-Indian Childhood Cirrhosis (NICC), a form of early
childhood liver cirrhosis. This epidemiological study examines the exposition of infants to increased copper
concentrations through drinking water from public water supplies in Berlin, Germany, and if this dietary copper
intake can cause liver damage in early childhood. In total, water samples from 2944 households with infants were
tested for copper. Mean copper concentrations in the two different types of collected composite samples were 0.44
and 0.56 mgyl, respectively. Families having a copper concentration at or above 0.8 mgyl in one or both of the
composite samples (29.9% of all sampled households) and a defined minimum ingestion of tap water of their infant
were recommended to undergo a paediatric examination. Nearly every of the 541 recommended infants were examined
by a local paediatrician and of these 183 received a blood serum analysis, too. None of the infants had clear signs of
a liver disease although a few serum parameters lay outside the accompanying reference range and abdominal
ultrasound imaging gave slightly unusual results in five cases. Additionally, no signs of a negative health effect could
be found in the statistical analysis of the serum parameters GOT, GPT, GGT, total bilirubin, serum copper, or
ceruloplasmin in relation to estimated daily and total copper intakes of the infants from tap water. No dose relation
of serum parameters and estimated copper intakes could be established. From the results of the study, no confirmed
indication of a liver malfunction in infants whose food had been prepared using tap water with an elevated copper
concentration could be found and, therefore, no indication of a hazard due to copper pipes connected to public water
supplies could be detected.
䊚 2002 Elsevier Science B.V. All rights reserved.
Keywords: Copper; Heavy metals; Non-Indian Childhood Cirrhosis, NICC; Public water supply; Tap water
0048-9697/03/$ - see front matter 䊚 2002 Elsevier Science B.V. All rights reserved.
PII: S 0 0 4 8 - 9 6 9 7 Ž 0 2 . 0 0 3 9 9 - 6
128 B.P. Zietz et al. / The Science of the Total Environment 302 (2003) 127–144
mately 8 to 10 months of age, all families willing For some households pH value, conductivity,
to participate were asked to collect two different total hardness, SO2y 1y 3y
4 -, NO3 -, PO4 - and F
1y
composite samples of drinking water in rinsed concentrations were determined from an additional
150-ml PE bottles sent to them by post. Composite flushed water sample (water was allowed to run
sample type 1 was collected by taking an aliquot from the tap until there was a noticeable difference
of 100 ml of tap water each time it was used in in its temperature, then 500 ml was collected). Ion
the household. On the next day, collection of 250 chromatography was used to analyse the water for
ml of tap water in the morning, at noon, in the SO2y 1y 3y
4 -, NO3 -, PO4 - and F-concentrations. A
evening, and before going to bed was required for detailed description of used analysis methods can
composite sample type 2 (general course of the be found in Zietz et al. (2001).
day). The collected water for both sample types A family with copper concentrations of 0.8 mgy
had to be stored during the day in a clean non- l or higher in one of the composite samples was
metal container of the household until the PE recommended to have a paediatric examination if
transport bottles were filled with the composite their infant had received 200 ml of tap water per
samples in the evening. The families were instruct- day or more for at least 6 weeks. The copper limit
ed to cover the container between each collection. of 0.8 mgyl has been chosen for this study because
In the initial phase of the study, only one screening this value nearly equals the average copper content
sample taken by the collection of a small aliquot in human milk (0.72 mgykg; Souci et al., 1994).
of tap water filled directly into the 150-ml PE If the copper concentration exceeded 2.0 mgyl
bottles each time tap water was used for baby food (limit of the WHO), a paediatric examination was
preparation was tested. Additional composite sam- recommended even for exposures below 200 ml
ples type 1 and 2 were collected if 0.5 mg copper per day. All other families with elevated copper
per liter were reached or exceeded in the house- concentrations in their composite samples were not
hold’s screening samples and the infant was not recommended to have a paediatric examination of
fully breastfed. their infant. Each of the paediatricians involved
For all households, a questionnaire served to was notified about the background and aim of the
asses factors of interest regarding installation, study. The paediatric examination was recommend-
building and uptake of tap water. Tap water sam- ed to include the drawing of a blood sample and
ples were collected between June 1998 and March an examination of the liver using palpation and
2001. Due to practical reasons and remaining ultrasound imaging. During the blood serum anal-
uncertainties in food copper concentrations, a ysis the following parameters were determined:
detailed feeding history was not recorded. A socio- serum copper (colorimetic assay with PAESA,
economic analysis of the families has not been Nobis, Endingen, Germany); ceruloplasmin (neph-
made. elometry with rabbit antiserum, Behring, Marburg,
Following collection in rinsed 150-ml PE bot- Germany); immunoglobulins (IgG, IgM, IgA; tur-
tles, samples were acidified at pH-0.2 with nitric bidimetry with rabbit antisera, Behring); glutamate
acid. The copper concentration of the water sam- oxaloacetate transaminase (GOT); glutamic pyru-
ples was measured by atomic absorption spec- vic transaminase (GPT); gamma-glutamyl tran-
trometry using a graphite furnace Perkin–Elmer speptidase (GGT) (enzymes measured with
SIMAA 6000 following German industrial stan- photometric assay, Beckman Coulter, Galway, Ire-
dard DIN 38 406-7 (wavelength: 324.8 nm, detec- land); total bilirubin (colorimetric assay with diazo
tion limit: 0.005 mg Cuyl). For internal quality reagent, Beckman Coulter); and CRP (turbidime-
control, the reference material Perkin–Elmer Qual- try; Beckman Coulter). Immunoglobulins and CRP
ity Control Standard-21 elements PE pure was were measured to detect signs of infections that
used. The laboratory has successfully participated may alter the other serum parameters. For ethical
in regular external quality assurance for all meas- and practical reasons a control group for serum
ured parameters. values was not suitable.
130 B.P. Zietz et al. / The Science of the Total Environment 302 (2003) 127–144
Table 2
Chemical and physical parameters of selected households located in different suburbs of Berlina
have a copper installation, this relationship was that at least for this parameter there was no
also clearly visible and measured concentrations significant difference in the motivation of families
fell until the age class of 40–45 years. There was to participate in the study.
a slightly irregular decline and increase of mean
concentrations compared to values of all house- 3.2. Health status of exposed infants
holds (data not shown).
Gender relation of the last born infant stated by In total, a paediatric examination was recom-
participating families was very close to 1 to 1 mended for 541 infants. Nine families fulfilling
(excluding twins and triple births). This indicates the requirements for a recommendation of a pae-
Fig. 1. Mean copper concentrations and stated age of installation in composite sample type 1 and 2 (excluding samples collected
in relation to screening samples).
132 B.P. Zietz et al. / The Science of the Total Environment 302 (2003) 127–144
Fig. 2. Frequencies of all measured infant values of serum GOT, GPT and GGT grouped in classes.
diatric examination could not be contacted any- One blood serum analysis of an infant (ID No.
more. Nearly every of these exposed infants were B4804) was excluded from statistics and depiction
inspected and examined by a physician; in many in the tables because by mistake a full-blood
cases during regular examinations recommended sample was sent to our laboratory showing severe
for every infant and child in Germany. Only in six haemolysis. A second blood sample was refused
cases was a medical examination completely by the mother. Physical examination and ultra-
refused. In one case, an examination was not done sound imaging of the liver gave no indication of
because the paediatrician had found that there was a liver disease. The child had received an estimated
actually no exposition of the infant. In eight other 800 ml of tap water daily beginning at the 35th
cases, a result of a possible medical examination week (copper concentrations in tap water: com-
could not be acquired. posite sample type 1: 0.85 mgyl; and type 2: 1.05
None of the examined infants showed any symp- mgyl). Additionally a second case was also
tom of liver damage such as icterus or frequent excluded from statistics and depiction in the tables.
vomiting. There was no clinical suspect of a This infant (ID No. B4759) was known to have a
paediatrician that a child is suffering form a liver congenital infection with Toxoplasma gondii
disease. Results of inspection and liver palpation including an affection of the liver and was treated
were only evaluated by the local paediatricians with chemotherapeutics. In the frequent controls
(no external standardisation). In total, 183 infants of this infant, transaminases and cholinesterase
and children received a paediatric examination were slightly or moderately elevated in many
including a blood serum analysis (Fig. 2). Of these cases. This infant also had an increased liver
blood samples, 124 were analysed in the laboratory volume. The copper concentrations in this house-
¨
of the University of Gottingen. In one of these hold were 1.15 mgyl in composite sample type 1
examined infants, the copper value in a composite and 1.6 mgyl in type 2. The infant had received
sample of the household exceeded 3 mgyl and in 1000 ml of this tap water per day beginning at its
another case equalled 3.0 mgyl). Of 183 infants 19th week of life.
examined with a blood serum analysis, eight had There were eight infants with outlying values of
never been breastfed and in total 34 had received GOT, GPT, GGT, or serum copper and no outlier
drinking water beginning with their birth. of total bilirubin. Values were defined as outliers
B.P. Zietz et al. / The Science of the Total Environment 302 (2003) 127–144 133
if the mean of this serum parameter was exceed ing. Additionally, bile ducts in the liver were
by four standard deviations or if it had a distance slightly rarefied. Serum values of copper, GOT,
of the mean with a chance of occurrence below GPT, GGT were in the reference ranges and the
Ps0.05. Details of these infants can be found in total bilirubin value (1.0 mgydl) was at the upper
Table 3. Six of these infants had clinically diag- reference limit (Witt and Trendelenburg, 1982).
nosed infections at the time of the serum analysis Three other infants were diagnosed to have a slight
and one was known to have a liver haemangioma splenomegaly. Two of these three cases of slight
and a ureteric obstruction including hydronephros- splenomegaly are clearly associated with an infec-
is. Only in one case, there was no hint of a disease tive disease. It is very probable that the infection
that may be associated with the outlying serum was the reason for the (transient) splenomegaly in
parameter. Although GOT, GPT, GGT are indica- two of these cases. Infections are a known cause
tors of liver or muscle diseases, it seems likely of elevated spleen size (O’Reilly, 1998), due to its
that infections attribute to these minor elevations immunologic function. These are, therefore, the
of the parameters. The liver and the muscles can most likely cause of these slightly abnormal results
be affected during infection and using medication in these cases.
during infection can also be a cause of elevations
(Pratt and Kaplan, 2000). Serum copper is typi- 3.3. Statistics using results of serum analysis
cally be elevated by infective diseases (Beshgetoor
and Hambidge, 1998). None of these outliers were Based on the Kolmogorov–Smirnov test only
excessively elevated above the used reference val- serum copper and IgG were normally distributed
ues. Three of these infants had received a serum among all measured serum values of exposed
control analysis (B990, B2437a, B2437b). In all infants. Different daily and total copper intakes
cases, these infants had an infection at the time of through tap water were calculated for an examined
the first blood sample. Two of these (B2437a, infant by separately using copper concentrations
B2437b) had much lower values in the serum of composite sample type 1 and 2 and a mean of
controls. One infant (B990) did not have normal both types combined with parent’s statements of
values of GOT, but it was infected again at the volume and duration of tap water ingestion. The
time of the control sample. Details of other infants different estimated daily and total copper intakes
that received a serum control were given in Table were not normally distributed. Therefore, the
3, too. Spearman rank correlation between estimated
In five cases (slightly) unusual ultrasound imag- intakes and serum parameters was calculated (one-
es of the liver or spleen were found (serum sided and two-sided significance). No significant
parameters and a detailed exposition history are correlation of GOT, GPT, GGT, total bilirubin,
also given in Table 3). Only elevations of serum serum copper, or ceruloplasmin and estimated
copper (no definitive reference value available; copper intake through tap water could be found
Kirsten et al., 1985), total bilirubin, GOT, GPT, except one. Assuming a one-sided significance
GGT and ceruloplasmin were mentioned in the (P-0.05; prerequisite: direction of the effect is
following (Witt and Trendelenburg, 1982; Thomas, known) in one case there was a significant rank
1984). None of the examined infants showed any correlation of y0.133 between serum copper and
symptom of liver damage such as nausea or vom- daily copper intake calculated with composite sam-
iting. An 11-month-old girl had a slight hepato- ple type 1. This correlation could not be found
megaly (ID B802) with a liver length of 9.5 cm with all other estimated daily and total copper
that is 0.5 cm above the reference value of this intakes or using a two-sided significance Fig. 3.
body length class. Serum copper, total bilirubin, In a next step, complete records of infants
GOT, GPT, GGT and ceruloplasmin were in the including outliers in one serum parameter that
reference ranges. In another case, an 11-month-old exceed the mean by four standard deviations or
girl (ID B799) showed a slightly enhanced echo- that had a distance of the mean with a chance of
genicity throughout the liver with ultrasound imag- occurrence below Ps0.05 were eliminated. No
134
Table 3
Results of blood serum analyses of infants exposed to elevated copper concentrations having outliers in their serum parameters (serum copper, GOT, GPT, GGT) or
unusual ultrasound images of the liver or spleen or serum analyses that were repeated
B.P. Zietz et al. / The Science of the Total Environment 302 (2003) 127–144
B.P. Zietz et al. / The Science of the Total Environment 302 (2003) 127–144 135
Table 3 (Continued)
136
Table 3 (Continued)
B.P. Zietz et al. / The Science of the Total Environment 302 (2003) 127–144
Used reference values: GOT 2nd–12th month -28 mUyml, )12 months -23 mUyml (Witt and Trendelenburg, 1982); GPT 2nd–12th month -31 mUyml, )
12 months -25 mUyml (Witt and Trendelenburg, 1982); GGT 2nd–12th month up to 100 mUyml, )12 months -21 mUyml (Witt and Trendelenburg, 1982); total
bilirubin 2 months and older: up to 1 mgydl (Witt and Trendelenburg, 1982); Ceruloplasmin 15–60 mgydl (Thomas, 1984); IgM 7th–12th month 0.36–1.04 gyl, 2
years 0.72–1.60 gyl (Uffelmann et al., 1970); IgG 7th–12th month 3.5–11.8 gyl, 2 years 5.2–10.8 gyl (Uffelmann et al., 1970); IgA 7th–12th month and 2 years
0.36–1.65 gyl (Uffelmann et al., 1970); CRP up to 1 mgydl (Sabel and Wadsworth, 1979). Serum copper: 6th–12th month mean 136–161 mgydl, standard deviation
24–32 mgydl (Kirsten et al., 1985). n.t.sNot tested; B2437a and B2437b are twins.
a
Blood serum tested in foreign laboratory.
B.P. Zietz et al. / The Science of the Total Environment 302 (2003) 127–144 137
Fig. 3. (a) Serum GOT and estimated total copper intake through tap water (calculated with copper concentrations from composite
sample type 1) including linear trend line (Excel䉸 ). Spearman rank correlation y0.090 with Ps0,228 (2-tailed). (b) Serum GOT
and estimated daily copper intake through tap water (calculated with copper concentrations from composite sample type 1) including
linear trend line. Spearman rank correlation y0.060 with Ps0.422 (2-tailed). (c) Serum copper and estimated total copper intake
through tap water (calculated with copper concentrations from composite sample type 1) including linear trend line. Spearman rank
correlation y0.071 with Ps0.357 (2-tailed). (d) Total bilirubin and estimated total copper intake through tap water (calculated
with copper concentrations from composite sample type 1) including linear trend line. Spearman rank correlation 0.101 with Ps
0.201 (2-tailed).
significant correlation of GOT, GPT, GGT, total concentrations were evaluated separately, too. This
bilirubin, serum copper, or ceruloplasmin and esti- was done because some parents may have stopped
mated copper intake (daily and total intakes) feeding their infants with this tap water and a
through tap water could be found after eliminating possible health effect may only be effective for a
these records. The same statement is valid if only limited time. In these groups, several significant
outliers of GOT, GPT, GGT, serum copper, or correlations could be seen (Spearman rank; P-
ceruloplasmin were eliminated (there were no 0.05; one-sided significance) of GOT, GPT, GGT,
outliers of total bilirubin). serum copper, or total bilirubin and different daily
Additionally data of male (ns102) and female and total copper intakes through tap water. In all
(ns81) infants were evaluated separately. Also cases, these correlations were weak ()0.30) and
subgroups including only infants whose blood in most cases they had a negative sign. This would
samples were taken not later than 14 days (ns mean that higher copper intake would lead to
53) or in another group 21 days (ns81) after lower serum parameters which is the opposite of
notification about the elevated copper tap water the expected effect. In the subgroup of male
138 B.P. Zietz et al. / The Science of the Total Environment 302 (2003) 127–144
mean absolute value of 6.53 mmolyl (median 6.62 (1999). The median copper concentration was 0.72
mmolyl). Expressed as percentage of total serum mgyl and the 90th percentile 2.11 mgyl. The
copper, there was a mean of 30.6% (median median concentration was 0.81 mgyl in Malmö
33.7%). The 90th percentile was 11.49 mmolyl for and 0.54 mgyl in Uppsala. Among the 430 children
the absolute value and 48.4% expressed as relative who participated in this part, the mean daily intake
value. Spearman rank correlation between different of water was 620"200 ml. An in-depth diary
estimated daily and total intakes (as described study investigated the effect of flushing on copper
above) and free copper values were calculated. No values. In this part, the median copper concentra-
significant correlation could be found. Beside a tions were 0.93 mgyl in unflushed drinking water,
correlation of serum copper or ceruloplasmin and 0.19 mgyl after flushing for 30 s and 0.08 mgyl
free copper values there were weak rank correla- after flushing until the water ran consistently cold.
tions ()0.22) between relative free copper values Falling median copper concentrations with increas-
and GPT or total bilirubin (P)0.025). Using the ing age of the installation (due to generation of a
U-test (Mann–Whitney) to compare the highest surface layer slowing down corrosion) could be
and lowest of three exposure groups under differ- proved in this study, too. Measured concentrations
ing circumstances as described above gave no in our study of the Berlin area fell with increasing
significant differences for free copper values. age of the water pipes until the age class of 35–
40 years and then increased again. This trend was
4. Discussion also clearly visible assessing only households that
explicitly stated to have a copper installation. The
Copper concentrations found in this study are rising concentrations in higher age classes may be
relatively high compared to that of other regions. caused by an increasing probability of repair and
A similar study of this institute in the area of partial replacement (or unknown total replace-
¨
Gottingen, Germany, showed lower copper concen- ments) of the installation in higher age classes.
trations in tap water (Dassel de Vergara et al.,
The comparison of the cited investigations points
1999 and unpublished results). The mean copper
out that (mean) copper concentrations in the user’s
¨
concentration in the Gottingen study was 0.18 mgy
tap water can vary in a broad range. The Berlin
l in the 1619 collected stagnation samples and
water supply facilities do not add anticorrosive
0.11 mgyl in the 1660 random daytime samples.
substances (for example based on a silicate or
In Western Germany in the years 1985–1986, a
mean copper concentration of 0.248 mgyl (median phosphate) to their production of water (personal
0.067 mgyl) was measured in stagnated water communication, Berliner Wasserbetriebe, 2001).
samples collected in sampling places throughout In documented cases of NICC (Dieter et al.,
the country in 2577 households (Krause et al., 1999; Eife et al., 1991; Schafer¨ ¨
and Schumann,
1991). In random daytime samples, a mean of ¨
1991; von Muhlendahl and Lange, 1994; Schim-
0.154 mgyl and a median of 0.04 mgyl was found. melpfennig et al., 1997; Walker-Smith and Blom-
They could observe the trend that rural areas had field, 1973) in which the liver disease was
lower copper values than urban areas ()100 000 supposedly caused through tap water, but where
inhabitants). In water samples drawn in water data could be gathered only retrospectively, the
supply facilities (before entering the distribution copper concentrations of the tap water were found
network, ns97) a mean of 0.013 mg copper per to lie between 1 and 3 mgyl and were occasionally
litre and a median of 0.005 mgyl was measured. as high as 26.4 mgyl. Copper seems, therefore, to
Similar results were found in the years 1990–1992 be involved in the aetiology of NICC. The con-
for the total area of Germany (Becker et al., 1997). centration range of a few milligrams of copper per
In Uppsala and Malmo, ¨ Sweden, concentrations liter water was reached in several cases in this
of unflushed water samples (equivalent in most investigation. For example 40 households had
cases to stagnated water samples) were analysed copper concentrations of 2 mgyl or higher in
for 1178 children by Pettersson and Rasmussen composite sample type 2.
140 B.P. Zietz et al. / The Science of the Total Environment 302 (2003) 127–144
Aminotransferase (GOT and GPT) levels are transferase levels and the amount of copper ingest-
sensitive indicators of liver-cell injury and are ed with drinking water or the copper concen-
helpful in recognising hepatocellular diseases such trations in water was found.
as hepatitis (Pratt and Kaplan, 2000). Liver func- It is recommended as a first step in the evalua-
tion indicators such as albumin and clotting factors tion of a patient with elevated liver enzyme levels,
(e.g. prothrombin time) are a component of grad- but no symptoms, to repeat the test to confirm the
ing systems for chronic liver diseases. They do result. If the result is still abnormal, the physician
not have a good performance as a screening test should evaluate the degree of the elevation. A
(Rochling, 2001). Both aminotransferases are nor- minor elevation (less than twice the normal value)
mally present in serum at low levels (Pratt and may be of no clinical importance if common
Kaplan, 2000). Because blood serum analyses of disorders have been ruled out and, in fact, may
infants examined in this study were also made in not even be abnormal (Pratt and Kaplan, 2000).
several other laboratories in Germany, we know While working on the study, we used reference
that reference values of aminotransferases vary values based on several studies (Witt and Trende-
widely among laboratories ()25%). This is also lenburg, 1982; Sabel and Wadsworth, 1979; Uff-
described in the literature (Pratt and Kaplan, elmann et al., 1970; Thomas, 1984). A repeat of
2000). The normal range for any laboratory test is the serum analysis was recommended to the phy-
the mean value in a group of healthy persons "2 sicians if two parameters possibly relevant to liver
standard deviations. Thus, 5% of the results diseases (GOT, GPT, GGT, total bilirubin, serum
obtained from these group can be expected to be copper, ceruloplasmin) were outside the reference
outside the defined normal range and 2.5% above range or one parameter was elevated more than
the upper limit of normal (Pratt and Kaplan, 2000). 20% above the reference limit. In many cases, the
Data to define reference limits in infants and paediatricians having more information about these
children are often limited in number of available infants including the results of their clinical exam-
studies and number of examined persons. For ination did not confirm this recommendation. Gen-
example, used reference values of GOT, GPT and erally, the following hepatic causes of chronically
GGT are based on analyses of 210–254 infants elevated aminotransferase levels (GOT, GPT) are
and children measured in 15 different paediatric known in adults: alcohol abuse; medication; chron-
hospitals (Witt and Trendelenburg, 1982). It can ic hepatitis B and C; steatosis and non-alcoholic
be seen that our study has tested nearly as much steatohepatitis; autoimmune hepatitis; hemochro-
infants and children (ns183). This fact is also matosis; Wilson’s disease (in patients F40 years
true for C-reactive protein (CRP), immunoglobu- old); and alpha1-antitrypsin deficiency (Pratt and
lins (Sabel and Wadsworth, 1979; Uffelmann et Kaplan, 2000). Non-hepatic causes are celiac
al., 1970) and serum copper (no definitive refer- sprue, inherited disorders of muscle metabolism,
ence values; Kirsten et al., 1985). This is in many acquired muscle diseases and strenuous exercise
cases aggravated by the fact that measured values (Pratt and Kaplan, 2000). Some of these diseases
in infants and children vary substantially with the can be excluded in infants and other possible
age (Kirsten et al., 1985). Because of these reasons diagnoses such as NICC have to be added.
we do not discuss every excess of a reference limit Slightly unusual ultrasound images were found
in detail. Nearly all of these elevations are mar- in two cases of the liver (slight hepatomegaly;
ginal. slightly enhanced echogenicity and bile ducts
We, therefore, conclude that the higher than slightly rarefied) and in three cases of the spleen.
expected number of infants with slightly elevated Ultrasound of the liver is used as a screening
aminotransferase levels referring to cited literature imaging tool in cases of patients suspicious for
does probably not reflect a pathologic status of the diffuse liver disease and is helpful in the term of
tested infants. This assessment is confirmed by the follow-up examinations (Hammerstingl et al.,
fact that no positive correlation between amino- 2001). In the case of liver cirrhosis, there are
B.P. Zietz et al. / The Science of the Total Environment 302 (2003) 127–144 141
variable signs of homogenicity and echogenicity days, or in a second group no later than 21 days,
changes. Hepatomegaly and enhanced echogenici- no positive correlation of daily and total copper
ty can be signs of a pre-stage or an early stage of intakes and serum parameters could be found. In
liver cirrhosis. On the other hand, patterns of all cases, an infant was examined including a
enhanced echogenicity are unspecific and can be serum analysis, and we tried to contact the families
caused by different disorders of toxic agents (Ham- again to get information whether tap water was
merstingl et al., 2001). In case of NICC, the used for the infant after the notification about the
pathologic pattern of a micro-nodular cirrhosis is elevated copper concentrations. Most of the
to be expected, but in early stages a more subtle responding parents stated that they further used
and rather unspecific liver histopathology may be the tap water for their infants but many of them
present (Muller
¨ et al., 1998). In all cases of NICC flushed the water before using it for their infants.
for which data are available, biochemical markers It is known that flushing water leads to lower
of hepatic injury such as aminotransferases, alka- coppers concentrations (Pettersson and Rasmus-
line phosphatase, bilirubin, albumin and prothrom- sen, 1999). So it can be assumed that copper
bin time have been shown to be abnormal (Muller ¨ intake of the infants trend to be lower than in the
et al., 1998). In both described cases of abnormal weeks before notifying parents, but the detailed
liver images in this study, serum analysis did not amount of this copper intake remains unknown.
reveal any relevant excess of reference values and So there remains an uncertainty in estimating the
no clinical symptoms were seen. Three other copper intakes but using data of subgroups gives
infants were diagnosed to have a slight spleno- no hint that this would effect the results of the
megaly. In two of these three cases, it is very study.
probable that clinically diagnosed infections were We can summarise from this study that from the
the reasons for the slight splenomegaly. Spleno- paediatric examination (physical examination and
megaly is known to be a concomitant symptom of blood test) of the infants whose food had been
liver cirrhosis, but as a detached symptom it does prepared using tap water with a copper concentra-
not give proof of liver disease, because several tion of 0.8 mgyl or more (measured maximum 4.2
other systemic diseases such as haematological, mgyl), no confirmed indication of a liver malfunc-
infectious, congestive, or inflammatory disorders tion could be found. None of the examined infants
are known reasons, too (O’Reilly, 1998). Addi- showed symptoms of liver damage. Additionally
tionally, reference values in ultrasonic imaging of no negative health effects could be found in the
infants and children are only based on small statistical analysis of the serum parameters serum
examination groups similar to the situation in copper, GOT, GPT, GGT, total bilirubin, or ceru-
laboratory medicine, and results are influenced to loplasmin in relation to different estimated daily
some extent by personal experience and individual and total copper intakes of the infants from tap
interpretations of images (Konus et al., 1998; water.
Rosenberg et al., 1991; Bowie, 2000). Therefore, In an experimental study by Olivares et al.
we conclude that these slightly abnormal results (1998), no signs of disease were found in a group
in ultrasound imaging are, with a high probability, of infants whose baby food had been prepared
not a sign of liver disorder. with water having copper concentrations of 2 mgy
We could find no supporting facts concerning l. But it has to be taken into account that the
the hypothesis that an existing health effect may infants in the study of Olivares et al. (1998) were
have disappeared because many parents could have breastfed until they reached 3 months of age. Only
stopped feeding their infants with copper contam- at that time, exposition of the infant with copper
inated tap water and that, secondly, a possible supplemented drinking water started. This is in
health effect may only be effective for a limited contrast to many observed cases of NICC, that
time. Assessing only infants statistically whose were breastfed for a short time only or not breast-
serum samples were examined no later than 14 fed at all (Dieter et al., 1999). In our study of 183
142 B.P. Zietz et al. / The Science of the Total Environment 302 (2003) 127–144
¨
Eife R, Weiss M, Muller-Hocker¨ J, Lang T, Barros V, Sigmund Pratt DS, Kaplan MM. Evaluation of abnormal liver-enzyme
B, Thanner F, Welling P, Lange H, Wolf W, Rodeck B, results in asymptomatic patients. N Engl J Med
Kittel J, Schramel P, Reiter K. Chronic poisoning by copper 2000;342:1266 –1271.
in tap water: II. Copper intoxications with predominantly Rochling FA. Evaluation of abnormal liver tests. Clin Corner-
systemic symptoms. Eur J Med Res 1999;4:224 –228. stone 2001;3:1 –12.
European Union. Guideline 98y83yEG, 3rd November 1998. Rosenberg HK, Markowitz RI, Kolberg H, Park C, Hubbard
Amtsbl L330 issued December 5th 1998, p. 32. A, Bellah RD. Normal splenic size in infants and children:
Fitzgerald DJ. Safety guidelines for copper in water. Am J sonographic measurements. AJR Am J Roentgenol
Clin Nutr 1998;67:1098S–1102S (Suppl). 1991;157:119 –121.
Hammerstingl RM, Schwarz WV, Schmitt E, Faust D, Dietrich Sabel KG, Wadsworth C. C-reactive protein (CRP) in early
C, Zeuzem S, Vogl TJ. Bildgebende Diagnostik der Leber- diagnosis of neonatal septicemia. Acta Paediatr Scand
zirrhose. Radiologe 2001;41:852 –867.
1979;68:825 –831.
Kirsten GF, Heese HDEV, De Villiers S, Dempster WS,
¨
Schafer SG, Schumann
¨ K. Zur Toxikologie des Kupfers.
Poccock F, Varkevisser H. Serum zinc and copper levels in
the 1st year of life. S Afr Med J 1985;67:414 –418. Bundesgesundheitsbl 1991;7:323 –327.
Konus OL, Ozdemir A, Akkaya A, Erbas G, Celik H, Isik S. Scheinberg IH, Sternlieb I. Is Non-Indian Childhood Cirrhosis
Normal liver, spleen, and kidney dimensions in neonates, caused by excess dietary copper? Lancet 1994;344:1002 –
infants and children: evaluation with sonography. AJR Am 1004.
J Roentgenol 1998;171:1693 –1698. Schimmelpfennig R, Dieter HH, Tabert M, Meyer E. Fruh- ¨
Krause C, Chutsch M, Henke M, Leiske M, Meyer E, Schulz kindliche Leberzirrhose (FKZ) und Kupferexposition uber ¨
C, Schwarz E, Wolter R. Umwelt-Survey. Vol. IIIb, Wohn- das Leitungswasser. Umweltmed Forsch Prax 1997;2:63 –
Innenraum: Trinkwasser. Deskription der Spurenelementge- 70.
halte im Haushalts- und Wasserwerks-Trinkwasser der Sen BC. Infantile cirrhosis. Indian Med Gaz 1887;22:338.
Bundesrepublik Deutschland 1985y86. WaBoLu-Hefte,
Souci SW, Fachmann W, Kraut H (founded). Food composition
BGA, Berlin, 1991.
and nutrition tables, 5th edition. Stuttgart: Medpharm, 1994.
¨
von Muhlendahl KE, Lange H. Copper and childhood cirrhosis.
Lancet 1994;344:1515 –1516. Tanner MS. Indian childhood cirrhosis. In: Meadow SR, editor.
¨
Muller ¨
T, Muller W, Feichtinger H. Ideopathic copper toxico- Recent Advances in Paediatrics, vol. 8. Edinburgh: Churchill
sis. Am J Clin Nutr 1998;67:1082S–1086S (Suppl). Livingstone, 1986. p. 103 –120.
¨
Muller ¨
T, Schafer HJ, Rodeck B, Haupt G, Koch H, Bosse H, Tanner MS. Role of copper in Indian childhood cirrhosis. Am
Welling P, Lange H, Krech R, Feist D, Muhlendahl ¨ KE, J Clin Nutr 1998;67:1074S–1081S (Suppl).
¨
Bramswig ¨
J, Feichtinger H, Muller W. Familial clustering Tanner MS, Portmann B, Mowat AP, Williams R, Pandit AN,
of infantile cirrhosis in Northern Germany: a clue to the Mills CF, Bremner I. Increased hepatic copper concentration
etiology of idiopathic copper toxicosis. J Pediatr in Indian childhood cirrhosis. Lancet I 1979;8128:1203 –
1999a;135:189 –196. 1205.
¨
Muller ¨
T, van de Sluis B, Muller W, Pearson P, Wijmenga C. Thomas L (ed.). Coeruloplasmin (Cp). In: Labor und Diag-
Non-Indian Childhood Cirrhosis. Eur J Med Res nose. Indikation und Bewertung von Laborbefunden fur ¨ die
1999b;4:293 –297. medizinische Diagnostik. 2. Aufl. Medizinische Verlagsge-
¨ ¨
Muller-Hocker J. Pathomorphology of the liver in exogenic sellschaft, Marburg, 1984. S. 529–532.
infantile copper intoxication in Germany. Eur J Med Res Trollmann R, Behrens R. Spate ¨ Manifestation einer Indian
1999;4:229 –232.
Childhood Cirrhosis aufgrund einer chronischen Kupferin-
¨ ¨
Muller-Hocker J, Meyer U, Wiebecke B, Hubner¨ G, Eife R,
¨
toxikation uber saures Leitungswasser. Gesundheitswes
Kellner M, Schramel P. Copper storage disease of the liver
and chronic dietary copper intoxication in two further 1998;60:376 –377.
German infants mimicking Indian childhood cirrhosis. Path TrinkwV 1990. Verordnung uber¨ ¨
Trinkwasser und uber Wasser
Res Pract 1988;183:39 –45. ¨ Lebensmittelbetriebe (Trinkwasserverordnung-TrinkwV)
fur
Olivares M, Pizarro F, Speisky H, Lonnerdal ¨ B, Uauy R. vom 5.12.1990, BGBl I, p. 2612; BGBl I, 1991, p. 227.
Copper in infant nutrition: Safety of World Health Organi- TrinkwV 2001. Verordnung uber ¨ ¨ von Wasser fur
die Qualitat ¨
zation Provisional Guideline Value for Copper Content of den menschlichen Gebrauch (Trinkwasserverordnung-
Drinking Water. J Pediatr Gastroenterol 1988;26:251 –257. TrinkwV 2001). BGBl I, no. 24, issued May 28th 2001, p.
O’Reilly RA. Splenomegaly in 2505 patients in a large 959.
university medical center from 1913 to 1995, 1913 to 1962: Uffelmann JA, Engelhard WE, Jolliff CR. Quantitation of
2056 patients. West J Med 1998;169:78 –87. immunoglobulins in normal children. Clin Chem Acta
Pettersson R, Rasmussen F. Daily intake of copper from 1970;28:185 –192.
drinking water among young children in Sweden. Environ Walker-Smith J, Blomfield J. Wilson’s disease or chronic
Health Persp 1999;107:441 –446. copper poisoning? Arch Dis Child 1973;48:476 –479.
144 B.P. Zietz et al. / The Science of the Total Environment 302 (2003) 127–144
Witt I, Trendelenburg C. Gemeinsame Studie zur Erstellung Zietz B, Dassel de Vergara J, Kevekordes S, Dunkelberg H.
von Richtwerten fur ¨
¨ klinisch-chemische Kenngroßen im Lead contamination in tap water of households with children
Kindesalter. J Clin Chem Clin Biochem 1982;20:235 –242. in Lower Saxony, Germany. Sci Total Environ 2001;275:19 –
World Health Organization. Guidelines for drinking-water 26.
quality. Volume 1, Recommendations. Second edition.
WHO, Geneva, 1993, pp 46 and 174.